Setup Guide

Getting started with TigerGraph MCP requires a working TigerGraph instance and a Python environment.

Prerequisites

• Python: Version 3.10, 3.11, or 3.12.
• TigerGraph: Version 4.1 or later (On-prem, Savanna, or Docker).
• Environment: A .env file containing your TigerGraph connection details and (optionally) an OpenAI API key.

Installation Options

Option 1: PyPI (Recommended)

The simplest way to install the official MCP server:

bashterminal
Copy
pip install tigergraph-mcp

Option 2: Build from Source

If you want to contribute or explore the DevLabs implementation:

bashterminal
Copy
git clone https://github.com/TigerGraph-DevLabs/tigergraphx
cd tigergraph-mcp
poetry install

Configuring VS Code (GitHub Copilot)

To use TigerGraph as a tool within VS Code Copilot Chat:

1. Create a .vscode/mcp.json file.
2. Define the TigerGraph MCP server path and environment variables.
3. Use the @mcp command in Copilot Chat to interact with your graph.

Community Implementation

For advanced features like the AG2 agent framework integration, you can use the Custom Discoveries repository:

bashterminal
Copy
git clone https://github.com/custom-discoveries/TigerGraph_MCP

[!IMPORTANT] Ensure your TigerGraph user has the necessary permissions (e.g., QUERY_READER or ADMIN) depending on the tools you wish to expose to the LLM.